Artificial Bee-Optimized CNN for Osteoporosis Detection using Leg X-ray Images

Osteoporosis is increasingly recognized as a worldwide health concern as people live longer. Although, Dual-Energy-X-Ray Absorptiometry (DXA) is one of the primary diagnosis techniques for diagnosing Osteoporosis, its widespread use as a screening tool is limited. This study uses an evolutionary-based deep learning algorithm to forecast osteoporosis using leg X-ray images. Further, the Convolutional Neural Network was enhanced using the artificial bee optimization algorithm to classify normal and osteoporosis cases using leg X-ray images. The effectiveness of the developed Bee CNN was compared to the conventional CNN. The results indicate that the CNN combined with bee optimization achieved a high accuracy rate of 92.32% in accurately classifying normal and osteoporosis X-ray leg images. Integrating artificial bee optimization into the CNN demonstrated superior performance compared to the conventional CNN. Based on these results, the Artificial Bee CNN model proves to be a valuable tool for straightforward osteoporosis prediction in practical clinical environments.